Two A.C. output signals amplitude-modulated in accordance with two function values (sine and cosine) differing from each other in correspondence to a position-to-be-detected are received from a position sensor such as a resolver. By performing an addition or subtraction between a signal derived by shifting the electric phase of one of the received A.C. output signals by a predetermined angle, and the other received signal, two electric A.C. signals (sin(&ohgr;t±d+&thgr;), sin(&ohgr;t±d−&thgr;)) are electrically synthesized which have electric phase angles (&thgr;) corresponding to the position-to-be-detected and are phase-shifted in opposite directions. “±d” here represents phase variation error caused by factors, other than the position-to-be-detected, such as temperature change. In the synthesized two signals, the phase variation errors (±d) appear in the same direction, while the phase differences (&thgr;) corresponding to the position are shifted in opposite, positive and negative, directions. Thus, by measuring the respective phase shift amounts (±d+&thgr;, ±d−&thgr;) and performing appropriate operation, it is allowed to cancel out or extract the error (±d) so that an accurate phase difference (&thgr;) can be detected. Position detection data indicative of the detected phase difference (&thgr;) is converted into a pulse-width-modulated signal and transmitted in the pulse-width-modulated form.